Tencent Holdings, the Chinese games and social-media group behind Weixin and WeChat, reported third-quarter revenue of 207.34 billion yuan on November 18, up 8% from a year earlier but below the 210.77 billion yuan analyst estimate cited in the report, using LSEG data. Growth slowed from 11% in the previous quarter as gaming momentum cooled, even as artificial intelligence demand lifted cloud services.
The Hong Kong-listed group is extending its established games, messaging, advertising, payments and cloud businesses into AI models and applications. Its distribution and earnings give it resources for that expansion. The immediate financial strain is substantial: operating profit fell 7%, while compute procurement helped turn a net cash position into net debt.
Tencent’s Q3 2026 revenue and net profit compared with analyst expectations. Graphic: Magica; data: Reuters via Channel NewsAsia. Source: Channel NewsAsia.
A revenue miss alongside a profit beat
Revenue fell about 3.43 billion yuan, or 1.6%, short of the reported consensus. Net profit attributable to shareholders nevertheless rose 0.8% to 63.61 billion yuan, exceeding the 59.46 billion yuan analyst expectation cited in the same report.
That headline profit increase masks weaker operating performance. Tencent’s results filing shows operating profit declining 7% to 59.15 billion yuan. Its non-IFRS adjusted net profit was 66.52 billion yuan, down 6% from 70.55 billion yuan a year earlier. Those supplementary measures exclude items including share-based compensation and certain investment-related effects.
Net gains from investments and other items increased to 10.06 billion yuan from 2.82 billion yuan, while income-tax expense fell 7%. Tencent attributed the tax reduction mainly to a particular overseas subsidiary; without that effect, tax expense would have risen. Meanwhile, its share of profits from associates and joint ventures declined. The small increase in reported net profit therefore does not establish that earnings from operations strengthened.
Binance Wallet says subscriptions for its Anthropic IPO xStocks campaign reached the 500 million USDC limit. Final allocations, token delivery and applicable refunds remain pending; the cap does not establish proceeds received by the AI company.
Editorial Team
Cloud accelerates as games face a tougher comparison
The filing shows how differently the main businesses performed in the three months ended September 30:
Business
Q3 2026 revenue
Change from Q3 2025
Domestic games
47.6 billion yuan
+11%
International games
18.9 billion yuan
−9%
Marketing services
43.0 billion yuan
+19%
FinTech and business services
63.7 billion yuan
+9%
Tencent attributed domestic games growth to titles including Delta Force, VALORANT MOBILE and Honour of Kings. It said Delta Force, released in 2024, and VALORANT MOBILE, released in 2025, each reached record daily active user accounts during the quarter. The international decline reflected a demanding comparison with excitement around games from Supercell, its Finnish games subsidiary, and sales of Dying Light: The Beast in the year-earlier quarter. Currency movements also mattered: international games revenue fell 4% at constant exchange rates, compared with the reported 9% decline.
Cloud services revenue rose 35%, which Tencent attributed to AI-related demand and international expansion. Tencent said it allocated more compute to selling access to models and renting graphics processing units, the chips used for AI computation. But cloud sits inside the broader FinTech and business services segment. The filing does not give a standalone cloud revenue figure, so that growth rate cannot be converted into a disclosed amount of AI revenue or compared directly with the group’s gaming sales.
The margin picture was mixed. The value-added services segment, which includes games and social-network services, lifted gross margin to 64% from 61%, helped by internally developed games. FinTech and business services held its gross margin at 50%, with cloud among the drivers of gross-profit growth. Group gross margin was broadly stable at 56% year on year, although down from 58% in the second quarter.
Tencent also has a large distribution base for AI features: its Weixin and WeChat messaging and social platforms had a combined 1.449 billion monthly active user accounts at September 30, up 2% over the year. It said improved AI advertising and content recommendations supported its marketing business. Yet marketing services’ gross margin fell to 52% from 57%, reflecting AI operating costs and content spending for its Mini Dramas service.
Technical infrastructure operating costs, excluding depreciation, more than doubled to 19.99 billion yuan from 8.62 billion yuan a year earlier. These are group-wide costs, rather than a disclosed AI-only bill.
General and administrative expenses rose 29% to 44.1 billion yuan, primarily because of research and development spending on Tencent’s Hy family of foundational AI models, Weixin AI initiatives and AI capabilities across its products. R&D expense reached 32.36 billion yuan, compared with 22.82 billion yuan a year earlier.
Tencent reported non-IFRS operating profit of 68.1 billion yuan. On the company’s calculation excluding contributions from new AI products, that measure would have been 83.3 billion yuan and grown 8% year on year, though it fell 3% from the second quarter. The roughly 15.2 billion yuan difference illustrates the current earnings burden of those products; it is not a standalone measure of total AI investment.
Capital expenditure reached 81.43 billion yuan—about 39% of quarterly revenue, calculated from the filing—compared with 12.98 billion yuan a year earlier and 52.78 billion yuan in the second quarter. This measure covers assets including IT infrastructure, data centres and other properties and intellectual property; it is broader than AI chip spending alone.
Cash payments follow a different timetable. Tencent reported 65.1 billion yuan of capital expenditure payments and substantial compute prepayments within operating cash flow. Quarterly free cash flow was negative 69.1 billion yuan. The company said it would have been positive 16.9 billion yuan without the compute prepayments, implying an 86 billion yuan cash-flow effect from those advances.
That distinction matters: advance payments create an immediate cash burden without representing an equivalent current-period operating expense. Tencent moved from net cash of 58.2 billion yuan at June 30 to net debt of 13.7 billion yuan at September 30, primarily because of large compute procurement, according to the filing.
Borrowings rose to 380.1 billion yuan from 298.9 billion yuan at June 30. Tencent still held 231.7 billion yuan in cash and cash equivalents and 294.0 billion yuan in term deposits and other treasury holdings at September 30; its net-debt measure offsets these against borrowings and notes payable.
Financing continued after the quarter. The filing discloses October issuance of $1.8 billion in dollar notes and 26 billion yuan in renminbi notes. Tencent said swaps effectively converted the dollar notes into renminbi debt at lower interest rates. The group also continued shareholder returns during the September quarter, repurchasing HK$7.0 billion of shares. These commitments make cash allocation relevant alongside the cost of the AI expansion.
Tencent said procurement supported its Hy models, inference—the computation needed to run models—for WorkBuddy, wider product development and external cloud demand. WorkBuddy is Tencent’s personal AI agent workspace. In its June 5 product announcement, the company had already introduced an enterprise version and described connecting workplace tools into that workspace.
Its November business update claimed WorkBuddy had exceeded 10 million daily active user accounts, collectively consuming more than 100 trillion model tokens a day. Those are company-provided usage measures, rather than disclosed paying-customer or revenue figures, and the update does not establish them as September-quarter averages.
Chronology also limits what the quarter can demonstrate. Tencent released its Hy4 model in November, after the September quarter had ended. Its latest model advances cannot yet be assessed through a full quarter of reported financial performance.
Supply remains another constraint. Nomura analysts cited in a contemporaneous report warned that tighter US restrictions on access to advanced AI computing could constrain Tencent’s infrastructure investment and slow model development and application launches. That is a risk assessment, rather than evidence of a quantified interruption in Tencent’s supplies.
The next financial test is whether faster cloud growth and greater AI usage produce enough additional gross profit to cover development and operating costs—and whether compute prepayments translate into capacity that earns a return. The filing does not isolate AI revenue, cloud profitability or revenue per WorkBuddy user. Paying-customer conversion, the cost of serving usage and the timing of further procurement payments will matter more to that assessment than token consumption alone.
Tencent says US restrictions on remote computing will likely slow its next AI model while leaving cloud revenue largely protected. Southeast Asian operators face licensing and financing uncertainty, though US cloud access and domestic demand offer partial alternatives.
Microsoft is rolling out Agent 365 Business Analytics through its Frontier early access program, pairing AI spending with estimates of assisted work. The dashboard expands its existing analytics as agent workloads add usage charges, but its dollar-value estimates do not establish financial returns.
Adobe’s Firefly update saves recurring visual subjects and creative preferences, extending an existing push into repeatable production. Memory has narrower reach than the assistant, while campaign demonstrations leave reliability and savings unmeasured.
Microsoft’s Ignite upgrades package enterprise knowledge, skills and tools into reusable business plugins. The expansion builds on IQ’s June release and competes with existing graph retrieval services; total savings and portability remain unproven.
The European Central Bank warns that debt and interlinked financing could carry an AI investment reversal into credit markets and the European economy. Resilient earnings, limited direct bank lending and stronger bank buffers temper its comparison with past bubbles.
Reported introductions helped Alice Huang secure Alibaba as a chip-rental customer and obtain financing. The relationship raises questions about Nvidia's overseas sales network, but does not itself establish illegal exports.
KPMG says AI agents pursued some issues more deeply in a parallel UK audit, though humans still had to correct their work. Client fee savings remain unconfirmed, even as KPMG International has negotiated lower fees from its own auditor by citing AI.
Payhawk announced a $280 million all-equity Series C led by Growth Equity at Goldman Sachs Alternatives, valuing it at $2 billion. The funding backs an expansion from managing business spending into broader AI finance operations, where competitors already combine automation, payments and controls.
A 16-country survey of 15- to 22-year-olds finds a median of 33% have used AI as a companion or friend. Users have higher adjusted odds of loneliness, but the research cannot establish whether AI helps, harms or replaces human connection.
South Korean President Lee Jae Myung has ordered a review of rewards of at least 100 million won for reporting technology-assisted exam cheating and legislation requiring technology companies to supply evidence. Ahead of the November 19 college entrance exam, the proposals would add informant incentives and post-exam investigations to classroom checks, but implementation rules remain unsettled.
Microsoft is pitching Copilot Workspaces as a shared environment for people and AI agents. Team collaboration is already in early access, but personal Autopilot integration is still coming soon, with final agent pricing unannounced.
Google is rolling out file references and sharing for Gemini skills as adult personal accounts lose the ability to use Gems. Automatic conversion preserves supported material, but file limits and missing tools can interrupt existing workflows.
Muddy Waters disclosed a short position in Nvidia while questioning a possible connection involving senior adviser Ajay K. Puri. Its own account leaves the key identity question unanswered, separating its new claims from an existing server-diversion prosecution.
Nvidia reported $115.8 billion in quarterly revenue in its first earnings release since co-founder Jensen Huang resigned. Strong demand gives interim CEO Colette Kress a financial cushion, while a campaign for Huang’s return collides with unresolved investigations and the board’s reported opposition.
OpenAI launched GPT-6.2 Astra and Sol for professional work, cutting Astra’s standard input and output token prices by 20%. The models offer cheaper paths through selected tasks, but research evaluations, reporting failures and unfinished independent safeguard testing complicate the case for broader autonomy.
Microsoft’s first live speech-to-speech model offers OpenAI API compatibility and planned enterprise integrations. Its published evaluations show a low-cost option, with several rivals completing more agentic tasks.
Canberra-based Trellis Data is offering its 220-billion-parameter Agentaus 1 model to business and government. Its pitch goes beyond keeping data in Australia, but performance, deployment costs and a planned capital raising remain key tests.
Perplexity is adding direct collaboration between frontier models to Computer, letting a lead agent coordinate up to 16 supporting agents. Company-reported benchmarks suggest mixed teams can outperform repeated copies of one model, but leave customer costs and production reliability unresolved.
Marsh reported 105 client cyber incidents in the last two weeks of October, against 48 in the final two weeks of May. The broker links the increase partly to AI, while warning that insurance panel access may not guarantee an available response team.